Butterfly Research at the University of Sussex
Originally published in the 2016 Autumn Newsletter
Perhaps the best way to start this article is by introducing myself. I moved to the University of Sussex 8 years ago, in February 2008, where I am Professor of Apiculture and head of the Laboratory of Apiculture and Social Insects (LASI) in the School of Life Sciences. As you would expect, most of my research in on social insects, with the honey bee being the main focus at the moment. But we also study, or have studied, ants, wasps, and stingless bees (in Brazil and Mexico).
For the past 6 years a major focus of LASI research has been on honey bee foraging and disease control, with the umbrella title of this project being The Sussex Plan for Honey Bee Health & Well Being. This research has led the LASI team into butterfly research. This is not surprising as both butterflies and bees visit flowers. This has led to research on flower visiting by butterflies.
From a personal perspective, it is also not surprising. As a boy, catching and collecting caterpillars, moths and butterflies was a hobby. I went to secondary school in West Sussex, at Christ’s Hospital near Horsham (1964-1970). In those days there were good numbers of Pearl-bordered and Small Pearl-bordered Fritillaries in the woods a few miles south of the school near Southwater, not to mention White Admirals and occasional Purple Emperors. Duke of Burgundies were also present. One of my friends at Christ’s Hospital was Matthew Oates, with whom I made many butterfly- catching trips. As far as I understand things, he still likes to chase Purple Emperors. For myself, I have never been back to the woods near Southwater, but should probably return one day. My own path in entomology took me to Dyce Laboratory for Honey Bee Studies at Cornell University in New York State to carry out my Masters and Doctoral degrees. I quickly became fascinated by honey bees, both for their amazing social behaviour and their importance to humans. I later also became fascinated by other social insects, including wasps and ants.
The attractiveness of garden flowers to insects
One area of research in the Sussex Plan has been on ornamental garden flowers. Many garden plants on sale in garden centres have logos on them, recommending them as being good for honey bees, or bees, or pollinators. One of these is the Royal Horticultural Society’s Perfect for Pollinators list. With PhD student Mihail Garbuzov we carried out several projects aimed at putting these recommendations on a firmer scientific footing.
In one project we set up a garden with 32 summer-flowering varieties, each in two 1x1m beds, on the university campus. We also set up two additional smaller gardens, each with 10 of these varieties; one in my own garden and one at Plumpton College. With the assistance of undergraduate student helpers we had hired, Mihail then spent two summers counting and identifying the insects seen. We found a huge variation, approximately 100- fold, among the 32 plant varieties in the numbers of insects attracted with the worst being Pelargonium, and the best being Origanum (marjoram) and Agastache.
The plant varieties all attracted more than one type of insect, but the different varieties attracted different mixes. Thus, Borage was mainly honey bees, versus mostly bumble bees for lavender and dahlia varieties. In further research, we found that honey bees will visit lavender in large numbers if the bumble bees are excluded. When bumble bees are present they out compete the honey bees, reducing nectar to low levels thereby making it not worthwhile for honey bees, which are slower than bumble bees and have shorter tongues, to visit the lavender. Among the plant varieties most attractive to butterflies were Origanum, Agastache, Everlasting Wallflower (Bowles Mauve), Hyssop and lavender. There was no plant variety on which butterflies predominated.
Bees were the most abundant insects seen, 84%, with bumble bees and honey bees predominating. After the bees, the hoverflies were the next most abundant at 9%. Butterflies and moths were only 2%. This is perhaps not surprising as, in my experience, butterflies are seldom very abundant in gardens compared to bees and hoverflies. However, I have seen that butterflies and moths can be the most numerous insects on flowers in chalk downland meadows in summer. For example, at Castle Hill 2km south of Falmer in East Sussex, or on the Albury Downs near Guildford in Surrey.
The butterflies we saw were from 16 species. All were common and widely-distributed species. The most numerous were the Meadow Brown, Small White and Small Tortoiseshell. There were also 4 moth species. The most often seen was the Hummingbird Hawkmoth, which was the third most seen Lepidoptera.
The attractiveness of different garden flower varieties to different butterfly species
This project followed on from the one above and was carried out in my garden in Magham Down, East Sussex, in the summer of 2013. The aim was to determine the mix of butterfly species visiting different summer-blooming garden flowers. To do this, I repeatedly counted the butterflies on 11 varieties in bloom in the garden. Most were herbaceous garden or wild flowers (Verbena bonariensis, Cat-mint, Hemp Agrimony, Purple Loosestrife, marjoram, teasel) or shrubs (Buddleia, Bowles Mauve, lavender) growing in flower beds. There was also bramble, growing in the hedge, and bird’s-foot trefoil growing in a meadow-lawn area. The flowers included the five recommended by Butterfly Conservation (BC) as being best for summer nectar (V. Bonariensis, Buddleia, Bowles Mauve, lavender and marjoram). The total area per variety varied from 1 to 15m2. In total, I recorded 2,659 visits from 14 butterfly and 1 moth species across 6 consecutive sunny days, 8th to 13th August, at the peak of garden butterfly activity.
With the help of LASI PhD student Kyle Shackleton, we collated and analysed the data. As the plants were growing in different-sized patches, some of which were in full bloom and others not, and as the period of data collection was relatively brief relative to the bloom period, the data were not used to determine relative attractiveness. Rather, we used the data to compare the mix of butterfly species seen on each plant. It turned out that this was very different, especially between Buddleia and marjoram. Buddleia attracted mainly the Large White and the brightly-coloured nymphalids (for example, Peacock and Painted Lady) while marjoram attracted many Meadow Browns, Gatekeepers, Small Whites, Green-veined Whites, and the few Small Coppers seen. The difference was quite striking: Gatekeeper 908 (marjoram) v. 0 (Buddleia); Meadow Brown 213 v. 6; Small White 141 v. 21; Green-veined White 27 v. 0; Small Copper 5 v. 0; Peacock 4 v. 312; Painted Lady 5 v. 123; Large White 2 v. 71. The most abundant species on lavender, Bowles Mauve, and also Cat-mint, was the Small White. The plant that had the highest biodiversity index of butterfly visitors was Hemp Agrimony; a species not in the five recommended by BC but a good one nevertheless, and also a native species common in Sussex.
Nectaring height preferences of butterflies
The project described above showed that Buddleia should probably not be called the ‘butterfly bush’. A better name would be ‘nymphalids + Large White bush’, although this is admittedly not a very catchy name. It also shows why it has such a reputation as being good for butterflies, as it attracts the large and brightly-coloured species commonly seen in gardens like the Peacock. Why was it not attractive to Meadow Browns and Gatekeepers? Kyle and I wondered if it was because Buddleia flowers are generally several metres above ground. Maybe some butterfly species do not like to nectar high up? Many butterflies are mainly found in meadow areas and these might well have a preference for nectaring low down.
We decided to investigate this in a follow-on project in which groups of four or five 10 litre pots of six varieties of garden flower, including marjoram and small Buddleia bushes, were placed on the ground, so that the flowers were low down, or elevated on posts, so that the flowers were c. 3m above ground. These were set up on the University of Sussex campus and in my garden. They attracted quite some attention from passers-by. Were we making some kind of oriental hanging garden? What were we doing!
After gathering data in the summers of 2014 and 2015, this project is nearly complete. After gathering of the final data in the summer of 2016, it will then be written up. What we did find was that many butterfly species are very reluctant to nectar at 3m, and these are the ones that were not seen on the Buddleia in the previous project. Even species that would nectar at 3m had a preference for 0.5m, and these were the species seen on Buddleia, such as the Large White and the nymphalids. By comparison, honey bees were seen in similar numbers both low and high. Honey bees probably cannot afford to have preferences. They have to go to where the food is in order to forage efficiently.
We also gathered information on the heights that butterflies flew at, both on the campus and in my garden. We followed thousands of flying individuals for 4 to 5 seconds and recorded whether or not they reached a height of 2m. These data were in good agreement with the other data. For example, Meadow Browns and Gatekeepers almost never reached a height of 2m, but Large Whites and nymphalids did so in about half of the observations. It seems, therefore, that some butterflies are ground- huggers and this affects the flowers they nectar on including whether they visit Buddleia.
It was hard to get enough data on many species despite many months of collecting data in two locations. For example, after the first two years we had hundreds of observations on common species, but on others not much at all: Silver-washed Fritillary 3 observations, Ringlet 2, Marbled White 2, Wall 4, Brimstone 1, Brown Argus 4, Small Copper 8, and Holly Blue 3.
One intriguing species is the Holly Blue, which seems to fly and nectar high up. This may be because it is a woodland species. But we need more data to be able to show this in a convincing way. Unfortunately, we don’t see as many Holly Blues as we would like!
Comparing flower constancy of bees and butterflies
One of the benefits of my job is that I get to travel to do research. For the past 15 years I have visited Brazil most winters. These visits are mainly to study ants and stingless bees, a group of social bees found in the tropics. Honey bees are abundant in Brazil but we rarely study them there as we have plenty in Sussex. Across the last two visits, in March 2015 and 2016, my PhD students (Hasan Al Toufailia, Nick Balfour, Kyle Shackelton) and I, with the help of a Brazilian colleague Denise Alves and several of her students, did a project in which we compared the flower constancy of bees and butterflies.
Flower constancy was first described by Aristotle over 2,000 years ago. If you follow a foraging honey bee you will see what he was referring to. The bee will almost certainly visit just a single type of flower, even if several species that bees visit are growing together. It is generally thought that bees are more constant than other insects. Follow a fly or butterfly and it is quite likely it will not be constant. I remember following one bee fly on the Sussex campus in spring. It visited 4 different flower species in 4 flower visits.
I had been looking for a natural setup where we could test this but I could never find what I was looking for in Sussex. But in Brazil, in the lawn in front of the entomology lecture building at the University of São Paulo in Piracibaba, we found a perfect study location. There were many small patches of two wildflowers, one yellow and one white, in bloom. Importantly, both were being visited by numerous insect species, including the honey bee, one species of stingless bee locally called Jatai ( Tetragonisca angustula), many butterfly species, and one wasp species. Four of the butterflies were common enough, and were visiting both plants, to allow us to study them. Our results showed that, as expected, the bees were more flower constant than the butterflies and the wasp.
It has been fun studying butterflies, especially those species that I chased as a schoolboy in Sussex and at home in Berkshire. The research we have done on butterflies has grown out of the research we are doing on bees. But that is often how things work in science, with one thing leading to another. One possible area of further research, combining both bee and butterfly angles, is to determine if butterflies and bees compete for nectar. This is a project I will be doing with Veronica Wignall, a new PhD student in LASI.
Another possible research topic could involve BC members in what is called ‘citizen science’. At the moment we are setting up a project in local parks in which we are planting 14 varieties of bee and insect-attractive plants, as well as 3 more that are not attractive for comparison. The idea is to help gardeners learn how to select the plants that are most beneficial for bees and other insects by learning to count and identify insects visiting flowers. A similar, but simpler project could be setup in the gardens of Butterfly Conservation – Sussex Branch members. The idea would be compare the mix of butterfly species visiting the five best for summer nectar plants recommended by BC, plus a few others such as Hemp Agrimony and Cat-mint. That is, which visit Buddleia, which marjoram, and so on. To make things more interesting, the Buddleia could be grown in two ways; in large bushes with flowers at 2-4m above ground and small bushes with the flowers at 0.5-1m.
Research publications discussed above
Garbuzov, M., Ratnieks, F.L.W. 2014. Quantifying variation among garden
plants in attractiveness to bees and other flower-visiting insects.
Functional Ecology, 28: 364-374. DOI: 10.1111/1365-2435.12178
Shackleton, K., Ratnieks, F.L.W. 2016. Garden varieties: How attractive are
recommended garden plants to butterflies? Journal of Insect Conservation
20: 141-148. DOI 10.1007/s10841-015-9827-9